【作者】 吴强辉；

【导师】 田学隆；

【作者基本信息】 重庆大学 ， 生物医学工程， 2005， 硕士

【摘要】 尿沉渣显微镜检查是进行临床检验和诊断鉴别的重要方法,是现代医学研究中常用的手段之一。但是,传统的光学显微镜观察是单人目视、主观观察,不但劳动强度大、容易引入误差,而且所看到的图像无法变换、处理,难以远距离传输,也不能进行快速、准确的定量处理,不能方便的存取,而数字图像处理技术恰恰能很好地解决这类问题。基于计算机系统的尿沉渣显微图像的处理和分析,除了极大地提高临床检验的效率,降低临床检验医生的劳动强度外,对医院的信息化、疾病诊断判别的标准化还将提供有益的帮助,也给医疗资源共享、远程会诊提供了便利。我们的工作就是研究开发一套专门针对尿沉渣显微图像分析处理的系统,使其具有更加完善的显微图像处理功能。本论文在理论和实践两方面,对尿沉渣有形成分的计算机处理和分析方法进行了深入的探讨。根据尿沉渣显微图像对象多、背景复杂、有许多干扰和噪声、目标深浅不一致、反衬度较小等特点,在对图像的各个处理过程(包括图像滤波、图像的边缘提取和边界跟踪、图像二值化、图像增强、图像分割、特征提取和特征分析等)的各种算法进行分析、比较的基础上,选择了对尿沉渣有形成分的分析有效的算法。其中图像分割是处理的重点,分割效果的好坏直接关系到图像分析的质量,我们在比较了各种分割方法的效果后,提出了基于类间方差最大法(即Otsu 法)的自适应阈值分割算法,从而有效地改善了尿沉渣显微图像分割的效果。为了尽可能完整地描述物体,本文在七个形态学参数对尿沉渣有形成分进行特征描述的基础上,另外从纹理学方面提取出能量、方差、熵、惯性矩、相关系数五个参数,以及一些光密度参数,给系统的进一步扩展奠定了基础。更多还原

【Abstract】 The urinary sediment inspection is an important methods of clinical diagnostication and one of the common means used in modern medical researching field . However, the traditional optic-microscopical inspection which is watched by one eye is easy to not only increase doctors’labour intensity but also bring up artificial error. Furthermore, the image can’t be transformed and processed by this means. It is very difficualty to remote-transmit, make precise measurement quickly and access conveniently the microscopical image. Whereas, all of this can be done by digital image processin technique. Processing and analysis of the urinary sediment’s micorscopical image based on computer system can increase efficiency of clinical inspection greatly, relieve the burden of doctors in clinical laboratory, provide standardization of disease diagnosis and manage present & history information of clinic. Moreover, it is easy to share medical datum and realize long-range consultation. The work we do is to researsch a analysis and processing system of urinary sediment microscopic image which will has greater power. The automatic processing and analysis of the visible image of urinary sediment have been discussed in detail on theoretical researching and practical realizing. It is complicated to process this kind of image which has several features such as more targets, complex background, much disturbance and noise, little contrast and so on. We compared and analysed all arithmetics on sediment image processing which includes filtrating, edge picking-up, boundary tracking, image enhancing, image segmenting, character picking-up, feature parameter selecting and so on. This paper put toward a set of effective algorithm to all kinds of urinary sediment. Among them image segmenting is very important that is directly relative to the precision of image analysis. After comparing the results of several segmenting arithmetics, we provide self-adaptation threshold value partition algorithm based on Otsu, which abviously improve the effect of urinary sediment segmenting. To integrallty describe sediment, this paper also extracts some illuminance density and five texture parameters such as energy, square-error, entropy, inertia matrix, correlation coefficient and so on, based on seven morphological parameters which can dicrib the feature of urinary sediment. All of this lay a foundation of the system function’s extending.更多还原